The MATIA Method™ Scale: the 5 AI maturity levels of an SME (and why none can be skipped)

Spectateur, Artisan, Orchestre, Architecte, Pionnier. Five metaphors to position your SME against AI in under ten minutes. And one golden rule that most business leaders try to sidestep, at their own cost.

Why yet another framework?

There is no shortage of AI maturity models. Gartner publishes a 5-level grid (Awareness → Transformational). Microsoft offers its Agentic L100 → L500 model. PwC talks of the AI-Native Enterprise. BCG, in its AI Radar 2026, distinguishes Followers / Pragmatists / Trailblazers. McKinsey and Deloitte refine their own frameworks every year. All these frameworks share one major flaw for a French SME business leader: they were designed for large corporates. They measure technological sophistication, model governance, data culture: concepts that do not carry the same meaning in an SME with 80 staff as they do at Saint-Gobain.

In practice, an SME boss who self-assesses with these grids systematically gets the same score: “level 1 out of 5”. The grid tells them to appoint a Chief AI Officer, set up an AI ethics committee, build an MLOps platform. None of these recommendations makes sense for their context. The business leader walks away frustrated and postpones the transformation.

On the French side, no official grid fills this gap. The 2025 France Num Barometer measures adoption as a percentage; Bpifrance's Diag Data IA works as an open-ended diagnostic with no formal level. The MATIA Method™ Scale fills this void. It was designed from more than thirty documented engagements in French SMEs and mid-caps since 2024. Five steps, each tied to a memorable metaphor and a simple question. The aim: that a business leader can position themselves in under ten minutes, understand where they are, where they need to go, and why they cannot skip the steps.

Overview: the 5 levels

Here is the full grid, with the estimated share of French SMEs at each level (cross-referencing Bpifrance Q2 2026, France Num 2025, KPMG Global AI Pulse Q1 2026, Stanford AI Index 2026 and BCG AI Radar 2026):

Total for levels 3+: ~16-18%, consistent with the Stanford AI Index 2026 (fewer than 10% have scaled AI within a function) and the share of “AI leaders” measured by KPMG in Q1 2026 (~11%). The consolidated picture: fewer than one French SME in five has crossed the threshold of measurable value; fewer than one in fifty is defining its sector.

The 9 dimensions of maturity, level by level

Beyond the overall positioning, the Scale reads across nine dimensions. Each one progresses from Spectateur to Pionnier; Level 5 describes the “golden standard 2030” target of the Pionnier, achieved today by fewer than 0.5% of SMEs, which is what makes it rare. The numerical thresholds (≈100%, < 1%, > 99.5%) are apex targets, not market averages.

Level 1: The Spectateur

“AI: we're watching from a distance. Are we missing something?”

The company is aware that generative AI exists. The business leader has heard about it at the chamber of commerce, in the business press, over a networking dinner. A few staff may have tried ChatGPT once, in a personal capacity. But no official tool is in place, no business use is identified, no budget line.

This level is not a moral failure: it is a state of watchfulness that has not turned into action. The danger is not being here in May 2026; it is still being here in May 2028. Meanwhile, competitors climb the steps, accumulate use cases and gain a few points of productivity every quarter. The gap widens silently.

Signs you are at Spectateur level

• No AI tool has been contracted or tested through a structured approach.
• The question “how many of our staff use AI?” has no answer.
• No AI budget is planned for the current financial year.
• The topic is handled by the executive committee as a matter to monitor, never as a matter for action.

Level 2: The Artisan

“Are my people using AI, each in their own corner?”

The company hosts individual, uncoordinated uses. The salesperson who writes their follow-ups with ChatGPT. The HR manager who pre-sorts applications with Claude. The technical director who trials Mistral. The business leader themselves who prepares their talks with an AI assistant. Each has subscribed to their own tool, sometimes on a personal account, without management knowing or formalising it. This is the territory of Shadow AI.

These uses are not a failure: they are a signal of appetite from the teams, a positive signal that must not be stifled. The problem is not that they exist; it is that they are private, fragile, invisible on the P&L, ungoverned. Three concrete risks characterise this level:

• Data exfiltration. A staff member pastes a client contract, an HR file, a pricing formula into a consumer assistant whose terms of use permit the data to be used to train the models.
• Undetected hallucinations. Without a usage framework, without training, users take incorrect outputs at face value (a wrong figure, an invented legal reference, a decision imagined in a set of minutes).
• Personal dependence. The know-how with the tool belongs to the staff member, not to the company. When they leave, they take their practice with them.

Signs you are at Artisan level

• You know your staff use AI, but neither how many nor exactly for what.
• No AI usage charter has been approved by the executive committee.
• No business use case is steered with a defined success indicator.
• No AI budget is a line in your P&L.
• No internal AI champion appointed.

Level 3: The Orchestre

“Is AI embedded in our processes, steered and measured?”

The company has identified 2 to 5 priority business use cases, deployed them with a genuine project approach, and measures their effects. An assistant for drafting quotes. A structured product-sheet generator. A customer follow-up tool. Each with an end user identified from the outset, success indicators defined before launch, a monthly review that goes up to the executive committee.

This level is not a final state. It is a state of governance. The company has realised that AI is a matter for executive management, not a matter for the IT department. It has appointed two to four internal champions, business operatives rather than technicians. It has written a usage charter that distinguishes what is permitted from what is not (notably on client data and intellectual property). It measures the benefits monthly, not straight off the production line but after reintegration into operational processes.

This is where the 159% median ROI documented on well-scoped AI engagements resides, according to Denis Atlan's SME AI-and-ROI barometer (159.8% over 24 months). 13 to 15% of French SMEs are here today, a cross-estimate between the share of “AI leaders” (≈11%, KPMG Global AI Pulse Q1 2026) and the 17% making regular use of generative AI (Bpifrance Le Lab, 82nd half-yearly barometer, January 2026).

Signs you are at Orchestre level

• 2 to 5 business use cases in production, measured.
• AI usage charter approved and communicated.
• Documented data policy (classification, access, retention).
• An AI committee at least quarterly.
• AI budget as a line in the P&L and ROI tracked.
• A register of AI systems maintained (AI Act preparation).

Level 4: The Architecte

“Is AI a structural competitive advantage?”

AI is no longer a project: it is an attribute of the business. The organisation has structured a proprietary knowledge base (typically 5 to 10 years of business archives cleaned, structured, indexed) which becomes the fuel for its AI agents. These agents are no longer mere assistants: they make operational decisions under human supervision, within defined scopes (quote generation with managerial validation, pre-processing of case files, drafting of visit reports).

The advantage is structural: a competitor wanting to replicate the capability would take 18 to 24 months: the time to build the equivalent knowledge base, clean it, structure it, and educate its teams to use it. This time gap becomes a measurable commercial differentiator: better response rates on tenders, higher quality of deliverables, greater speed of execution.

The Architecte is also characterised by documented governance: an AI Act register kept in real time (timeline updated after the Digital Omnibus of 7 May 2026), monthly AI committees, steered risk indicators (hallucinations, bias, data leaks, consumption), ISO 42001 initiated. This is the first level where AI FinOps becomes a mandatory discipline. Without an LLM gateway and without budget governance, costs spiral.

2 to 3% of French SMEs reach this level today. Reference BCG AI Radar 2026: 15% of “Trailblazer” leaders at the global level, adjusted downwards for the French SME; consistent with the ~16% of “Frontier” professionals who orchestrate agents (Microsoft, Work Trend Index 2026).

Signs you are at Architecte level

• A proprietary knowledge base in operation.
• Agents in production on at least one business scope, with documented supervision.
• AI Act register up to date, a monthly AI committee led by a member of the executive committee.
• An LLM gateway in place (routing, logging, cost control).
• An ISO 42001 policy being implemented.
• A data/AI architecture designed at enterprise scale.

Level 5: The Pionnier

“Are we setting the standard for our sector?”

The Pionnier is the company that no longer follows. It is followed. Its workflows are AI-first: the processes have been redesigned around the capabilities of agents, not the other way round. Its autonomous agents operate over durations measured in hours and then in days, under asynchronous supervision, with token budgets and contractual quality indicators. The company has certified ISO 42001 governance, a mature FinOps practice, a culture of “AI literacy” across leaders, managers and operatives, and the capacity to publish its own sector benchmarks.

Its “golden standard 2030” target reads across the nine dimensions above: mastered sovereign AI (RAG + targeted fine-tuning, local/on-premise option), splinternet-proof resilience (multi-cloud redundancy + on-premise failover), ~100% of staff augmented and ~100% cultural adoption, certified ISO 42001 governance, technical debt < 1% and a security score > 99.5%. These are apex thresholds: it is their conjunction, reached by fewer than 0.5% of SMEs, that sets the sector standard.

The Pionnier is not a large corporate. It is an SME or a mid-cap that made the early choice of deep integration and is reaping the competitive differential today. Microsoft, in its Work Trend Index 2026, talks of “Frontier Firms”. PwC talks of AI-Native. BCG, in its AI Radar 2026, talks of “Trailblazer” leaders. The vocabulary varies; the reality is the same: organisations that have made AI a constitutive attribute, not just one tool among others.

Fewer than 0.5% of French SMEs are at Pionnier level today. The Pionnier of 2026 will be the sector benchmark of 2030.

The golden rule: no level can be skipped

This rule is the most counter-intuitive and the most broken. An ambitious business leader at Artisan level sees an impressive demonstration of autonomous agents at a conference. They return to the office and launch a €200,000 project to build an agent platform for their company. Six months later, the project is halted. Not because the technology does not work: it works very well at Saint-Gobain. Because the foundations are not there.

In practice, skipping the Orchestre level means investing in a sophisticated platform while:

• the business data is neither structured nor accessible;
• the end users have no habit of working with AI;
• risk governance does not exist;
• no internal champion is able to carry the usage within their team;
• the success indicators were not defined before launch.

More than one time in two, this project fails. The CIO Playbook 2026 (IDC-Lenovo, January 2026) quantifies it: only 46% of POCs reach production. Half remain stuck at the pilot stage, the “industrialisation gap” that separates the minority genuinely transforming their operations from the majority left in permanent pilot mode.

Transition timescales and budgets (SMEs of 50-500 staff)

• Spectateur → Artisan: 3 to 6 months · €5,000 to €15,000
• Artisan → Orchestre: 6 to 12 months · €30,000 to €80,000
• Orchestre → Architecte: 12 to 24 months · €80,000 to €250,000
• Architecte → Pionnier: 24 to 36 months · a structuring investment, partly capitalisable off the P&L

A Spectateur → Architecte transformation takes 3 to 4 years. Spectateur → Pionnier takes 5 to 7 years. This is the irreducible duration of organisational maturation, independent of the speed of the technology. The tools accelerate; the people, the processes and the data follow their own pace.

How to position yourself in 10 minutes

Here is a quick test, derived from the 12-point MATIA Method™ grid used on engagements. Count your “yes” answers:

1. At least one business AI use case is in production with its indicators measured.
2. An AI usage charter is approved by the executive committee.
3. An internal AI champion (not from the IT department) is appointed and briefed.
4. A data classification policy is documented.
5. An AI budget is a line in the current year's P&L.
6. An AI committee meets at least quarterly with a member of the executive committee.
7. The AI Act register of AI systems in use is kept up to date.
8. Structured AI training has been delivered to more than 30% of the workforce.
9. A proprietary knowledge base is built or under construction.
10. At least one AI agent operates under supervision on a business process.
11. An LLM gateway (or equivalent) centralises calls and logging.
12. The ISO 42001 process is under way.

0 yes: Spectateur level. First objective: a 60-90 min diagnostic, a minimal charter, a first scoped use case.
1-2 yes: Artisan level. Structure the foundations (charter, champion, first use case with a project approach).
3-5 yes: Artisan → Orchestre transition. You have started; you must hold the discipline.
6-8 yes: stabilised Orchestre level. You are among the 13-15%. The first target for the 159% median ROI.
9-10 yes: Architecte level. A structural competitive advantage; FinOps and ISO 42001 become mandatory.
11-12 yes: Pionnier level. The standard for your sector. Capitalise and publish.

What is at stake for 2026-2028

The window is narrow. The leaders at Orchestre level by the end of 2025 will accumulate months of lead while the laggards hesitate. This gap of months will turn into a gap of years on acquisition costs, quality of deliverables, and execution capacity. The companies reaching Architecte level by the end of 2027 will hold a proprietary knowledge base that their competitors will not be able to replicate quickly.

The post-Digital Omnibus timeline (political agreement of 7 May 2026) offers a window (18 to 24 months) to install governance properly, before the high-risk AI Act obligations become binding at the end of 2027 / mid-2028. The SMEs that use this window to move Artisan → Orchestre, then Orchestre → Architecte, will gain a governance advantage that the laggards will only be able to catch up on in a rush: always more expensive, always more risky.

The good news: moving from Artisan to Orchestre is financially accessible to any SME or mid-cap of 50 to 500 staff. Total cost €30,000 to €80,000 over 12 months (support + licences + structuring). The only real investment is execution discipline.